On Mon, 16 Sep 2024 19:06:48 GMT, Kevin Driver <kdri...@openjdk.org> wrote:
>> src/java.base/share/classes/javax/crypto/KDF.java line 670: >> >>> 668: } >>> 669: } >>> 670: throw new InvalidAlgorithmParameterException( >> >> Should this ever happen? We already made sure `firstPairOfSpiAndProv` is for >> this algorithm and we have tried calling `deriveXyz` on it. Therefore there >> must be a `lastException` now. In the lines above, you checked two types of >> `lastException`. Is that exhaustive? If not, can you wrap it into an IAPE >> and throw it there? >> >> Also, even if `deriveXyz` finally fails, we still have to complete the DPS >> process, i.e. assign something to `pairOfSpiAndProv`, since DPS should only >> be run once. Later calls to `getProviderName` still need a return value. > > I may be wrong, but I think the final exception indicates a `Provider` > couldn't be chosen, so I'm not sure we can "complete" the provider selection > task. Maybe I misunderstood what you meant. My understanding is that there must be a non-null `lastException` when no provider is selected, and therefore you won't be able to reach here. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/20301#discussion_r1762033532
